Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. An electronic stabilization method, comprising: acquiring a frame of an image to be stabilized and a related exposure time; acquiring first attitude data before the exposure time and second attitude data after the exposure time, wherein the number of the first attitude data is one or more, and the number of the second attitude data is one or more; acquiring a frequency domain low-pass filter; acquiring a low frequency signal that does not exceed a cutoff frequency in the first attitude data and the second attitude data by filtering a high frequency signal that exceeds the cutoff frequency in the first attitude data and the second attitude data using the frequency domain low-pass filter; generating a target attitude corresponding to the exposure time based on the low frequency signal in the first attitude data and the second attitude data that does not exceed the cutoff frequency; and obtaining a stabilized target image by stabilizing the image to be stabilized according to the target attitude.
Image stabilization technology. This invention addresses the problem of stabilizing images captured by a device that experiences rotational movement during the exposure period. The method involves obtaining an image frame and its associated exposure time. Attitude data, representing the device's orientation, is collected both before and after the exposure time. This attitude data is then processed using a frequency domain low-pass filter. This filtering step removes high-frequency components from the attitude data, isolating low-frequency signals that represent slower, more significant rotational changes. A target attitude, representing the desired orientation for stabilization, is then calculated based on these filtered low-frequency attitude signals. Finally, the original image frame is stabilized to achieve the calculated target attitude, resulting in a stabilized target image.
2. The method of claim 1 , wherein before acquiring the frame of the image to be stabilized, the method further includes: acquiring a number of the frames of the image to be stabilized acquired by an image sensor of the image acquisition device, the image to be stabilized including a time stamp.
This invention relates to image stabilization techniques for improving the quality of images captured by an image acquisition device, such as a camera or smartphone. The problem addressed is the presence of unwanted motion or blur in captured images due to device movement, which degrades visual quality. The solution involves analyzing multiple frames of the image to be stabilized, each containing a time stamp, before acquiring the final frame. By processing these frames, the system can detect and compensate for motion, ensuring sharper and more stable output images. The method includes capturing a sequence of frames from an image sensor, where each frame is time-stamped to track its position in the sequence. This preprocessing step allows the system to assess motion patterns and apply stabilization algorithms to correct distortions before finalizing the image. The approach enhances image clarity by reducing motion artifacts, making it particularly useful in low-light conditions or when capturing fast-moving scenes. The invention improves upon existing stabilization methods by leveraging temporal data from multiple frames to achieve more accurate and efficient stabilization.
3. The method of claim 2 , wherein the image to be stabilized is stored in a first-in first-out memory.
A method for image stabilization involves processing a sequence of images to reduce motion-induced blur or distortion. The method includes storing the images in a first-in first-out (FIFO) memory buffer, which ensures that the oldest image is replaced by the newest one as the buffer reaches capacity. This storage approach maintains a continuous stream of recent images for real-time stabilization processing. The method further involves analyzing the stored images to detect motion vectors or other motion-related data, which are then used to apply corrective transformations to the images. These transformations may include spatial shifts, rotations, or scaling adjustments to compensate for detected motion. The stabilized images are then output for display or further processing. The FIFO memory ensures efficient handling of image data by preventing memory overflow while maintaining a sufficient history of images for accurate motion compensation. This technique is particularly useful in applications such as video recording, surveillance, and augmented reality, where smooth and stable visual output is critical. The method may also include additional steps such as filtering or interpolation to enhance stabilization quality.
4. The method of claim 2 , wherein before acquiring the frame of the image to be stabilized, the method further includes: acquiring attitude data obtained by a spatial attitude sensor of the image acquisition device, the attitude data including: the first attitude data and the second attitude data, or a third attitude data of the exposure time; the time stamp; and one attitude data that corresponds to the time stamp of each image to be stabilized.
This invention relates to image stabilization techniques for image acquisition devices, particularly addressing the challenge of compensating for unwanted motion or vibration during image capture. The method involves using spatial attitude sensors to detect the device's orientation and movement, ensuring sharper and more stable images. Before capturing a frame for stabilization, the method acquires attitude data from the device's spatial attitude sensor. This data includes multiple types of measurements: first and second attitude data, or third attitude data specific to the exposure time. Each set of attitude data is time-stamped to correlate with the exact moment of image capture. The method ensures that the attitude data accurately reflects the device's orientation during exposure, allowing for precise stabilization adjustments. The acquired attitude data is then used to analyze and compensate for motion artifacts, such as blur or distortion, in the captured image. By aligning the attitude data with the image's time stamp, the system can apply corrective measures to stabilize the final output. This approach enhances image quality by mitigating the effects of unintended movement, making it particularly useful in handheld or mobile imaging applications where stability is a concern. The method improves upon traditional stabilization techniques by incorporating multiple attitude measurements and precise time synchronization, ensuring more accurate and reliable results.
5. The method of claim 1 , wherein a first time period corresponding to the first attitude data and a second time period corresponding to the second attitude data are different.
This invention relates to systems and methods for processing attitude data of a moving object, such as an aircraft or spacecraft, to improve navigation accuracy. The problem addressed is the need to accurately determine the object's orientation over time, particularly when sensor data may be noisy or intermittent. The invention involves collecting first attitude data from a primary sensor during a first time period and second attitude data from a secondary sensor during a second time period, where the two time periods are different. The method then processes these datasets to refine the object's attitude estimation. The primary sensor may be a high-precision inertial measurement unit (IMU), while the secondary sensor could be a lower-precision but more robust system, such as a star tracker or magnetometer. By analyzing the overlapping or adjacent time periods of the two datasets, the method compensates for errors in one sensor using data from the other, improving overall accuracy. The invention is particularly useful in scenarios where sensor availability or reliability varies, such as during maneuvers or in environments with interference. The key innovation lies in the temporal separation of the datasets, allowing for cross-validation and error correction without requiring simultaneous operation of both sensors. This approach enhances navigation reliability in dynamic or challenging conditions.
6. The method of claim 5 , wherein a value range of the first time period and the second time period includes 0.5 second to 1 second.
This invention relates to a method for controlling a system, particularly focusing on timing parameters to optimize performance. The method involves adjusting a first time period and a second time period, where these periods define intervals for specific operations within the system. The key innovation is that the value range for both the first and second time periods is set between 0.5 seconds and 1 second. This range ensures that the system operates efficiently by balancing responsiveness and resource utilization. The method may be applied in various technical domains, such as industrial automation, communication systems, or data processing, where precise timing is critical. By restricting the time periods to this specific range, the system avoids delays that could degrade performance while preventing excessive resource consumption. The method may also include additional steps, such as monitoring system conditions and dynamically adjusting the time periods based on real-time data, to further enhance efficiency. The invention addresses the problem of optimizing system performance by providing a controlled and effective timing mechanism.
7. The method of claim 5 , wherein for two adjacent frames of image to be stabilized, the second attitude data of a previous frame of the image to be stabilized overlaps with the first attitude data of a subsequent frame of the image to be stabilized.
This invention relates to image stabilization techniques for video or sequential image frames, addressing the challenge of maintaining visual consistency between adjacent frames when capturing or processing images from a moving platform, such as a drone or handheld camera. The method involves using attitude data, which represents the orientation of the imaging device, to correct distortions caused by motion. The process begins by obtaining first attitude data for a subsequent frame and second attitude data for a previous frame. These data sets are then aligned such that the second attitude data of the previous frame overlaps with the first attitude data of the subsequent frame. This overlapping ensures a smooth transition between frames, reducing abrupt changes in orientation and improving stabilization accuracy. The overlapping data allows for interpolation or blending of motion corrections, minimizing artifacts like jitter or misalignment. The method may also involve calculating a transformation matrix based on the overlapping attitude data to apply corrections to the subsequent frame. This transformation compensates for rotational or positional shifts, ensuring that the stabilized output appears steady. The technique is particularly useful in applications where precise stabilization is critical, such as aerial videography, surveillance, or augmented reality. By leveraging overlapping attitude data, the method enhances the robustness of stabilization algorithms, providing smoother and more accurate results compared to non-overlapping approaches.
8. The method of claim 5 , wherein the image to be stabilized corresponds to a third time period, the third time period being less than a sum of the first time period and the second time period.
This invention relates to image stabilization techniques, specifically for stabilizing images captured over different time periods. The problem addressed is the need to stabilize an image that corresponds to a third time period, which is shorter than the combined duration of a first and a second time period. The method involves capturing multiple images over these time periods and processing them to achieve stabilization. The first time period corresponds to a first set of image data, while the second time period corresponds to a second set of image data. The third time period, being shorter than the sum of the first and second, allows for more precise stabilization by reducing motion blur and improving alignment. The method ensures that the stabilized image retains clarity and accuracy by leveraging the shorter third time period for processing. This approach is particularly useful in applications where rapid motion or unstable conditions require precise image stabilization, such as in surveillance, medical imaging, or high-speed photography. The technique optimizes the stabilization process by focusing on a reduced time frame, enhancing the overall quality of the stabilized output.
9. The method of claim 1 , wherein a delay between the exposure time of the target attitude and an exposure time of an actual attitude of the image acquisition device does not exceed a delay threshold.
This invention relates to image acquisition systems, particularly those used in dynamic environments where precise timing is critical. The problem addressed is the delay between capturing an image and the actual attitude (orientation and position) of the imaging device, which can lead to misalignment or blurring in applications requiring high accuracy, such as aerial surveillance, autonomous navigation, or satellite imaging. The method involves synchronizing the exposure timing of an image acquisition device with its target attitude to minimize delay. The system determines the target attitude the device should have at the moment of exposure and compares it to the actual attitude during exposure. If the delay between these two exceeds a predefined threshold, corrective measures are taken to ensure the image is captured within acceptable timing constraints. This may involve adjusting the exposure timing, compensating for motion, or recalculating the target attitude in real time. The invention ensures that images are captured when the device is in the correct orientation, reducing errors caused by motion or latency. This is particularly useful in scenarios where rapid changes in attitude occur, such as in drones, robotic cameras, or space-based imaging systems. The delay threshold is set based on the specific requirements of the application, balancing accuracy with operational constraints. The method may also integrate with other stabilization or tracking systems to further improve image quality.
10. The method of claim 9 , wherein a value range of the delay threshold includes 0-0.5 seconds.
Image processing and display systems are provided with a method for managing display updates to avoid visual artifacts or delays. The system aims to ensure smooth transitions and prevent the display of outdated information by determining an appropriate delay before presenting updated image data. A key aspect involves setting a threshold value that governs how long the system will wait, if necessary, before displaying a newly received image frame. This delay threshold is configurable, and in one specific embodiment, the allowable range for this delay threshold is set between 0 and 0.5 seconds. This allows the system to either display the new image immediately (0 seconds delay) or introduce a short delay up to half a second, balancing responsiveness with the need to avoid displaying partially rendered or incorrect frames.
11. The method of claim 1 , wherein a value range of the cutoff frequency of the frequency domain low-pass filter is between 0.5 Hz to 10 Hz.
This invention relates to signal processing, specifically to methods for filtering signals in the frequency domain to remove unwanted high-frequency noise while preserving low-frequency components. The problem addressed is the need to effectively filter signals with a tunable cutoff frequency to adapt to different noise conditions and signal characteristics. The method involves applying a frequency domain low-pass filter to an input signal, where the filter's cutoff frequency is adjustable within a specified range. The cutoff frequency determines the boundary between frequencies that are passed through and those that are attenuated. The invention specifies that the cutoff frequency of the filter should be set within a range of 0.5 Hz to 10 Hz. This range ensures that the filter effectively removes high-frequency noise while retaining the essential low-frequency components of the signal. The method may be applied to various types of signals, including but not limited to biomedical signals, environmental measurements, or industrial process signals, where noise reduction is critical for accurate analysis or control. The filter is implemented in the frequency domain, meaning the input signal is first transformed into its frequency representation, typically using a Fourier transform or similar technique. The low-pass filter is then applied to this frequency-domain representation, attenuating frequencies above the cutoff frequency. The filtered signal is subsequently transformed back into the time domain for further use. The adjustable cutoff frequency allows the method to be tailored to different applications, ensuring optimal noise suppression without distorting the signal of interest.
12. The method of claim 1 , wherein the frequency domain low-pass filter includes one or more of a FIR filter and an IIR filter.
Signal processing for audio or data transmission. The invention addresses issues in filtering signals in the frequency domain. Specifically, it describes a method where a frequency domain low-pass filter is employed. This low-pass filter is implemented using either a Finite Impulse Response (FIR) filter or an Infinite Impulse Response (IIR) filter, or a combination of both. This filtering process aims to attenuate high-frequency components within the signal while allowing lower frequencies to pass through. Such techniques are common in audio equalization, noise reduction, and signal conditioning to isolate desired frequency bands.
13. The method of claim 1 , further including: segmenting the image to be stabilized based on a segmentation method to obtain a plurality of sub-images; and stitching one or more of the plurality of sub-images based on the target attitude corresponding to the exposure time of the image to be stabilized to obtain a frame of stitched image, wherein the stitched image is the stabilized target image.
This invention relates to image stabilization techniques, specifically for improving the quality of images captured under unstable conditions, such as those taken by handheld devices or in dynamic environments. The problem addressed is the presence of motion blur or distortion in images due to camera movement during exposure, which degrades visual clarity and usability. The method involves segmenting an input image into multiple sub-images using a segmentation technique. Each sub-image represents a portion of the original image, allowing for localized processing. The segmented sub-images are then stitched together based on a target attitude corresponding to the exposure time of the original image. The target attitude refers to the desired orientation or position of the camera during exposure, which is used to align and combine the sub-images. The stitching process ensures that the final stitched image compensates for motion-induced distortions, resulting in a stabilized target image with reduced blur and improved clarity. The segmentation and stitching steps are performed to correct for motion artifacts, ensuring that the final output is a single, coherent image that appears as if it were captured under stable conditions. This approach enhances image quality by leveraging sub-image processing and precise alignment based on exposure timing and target attitude.
14. An image acquisition device comprising: a processor, an image sensor, and a spatial attitude sensor; wherein the processor is connected to the image sensor and the spatial attitude sensor; and wherein the processor is configured to: acquire a frame of an image to be stabilized and an exposure time of the image to be stabilized; acquire first attitude data before the exposure time and second attitude data after the exposure time, wherein the number of the first attitude data is one or more, and the number of the second attitude data is one or more; acquire a frequency domain low-pass filter; acquire a low frequency signal that does not exceed a cutoff frequency in the first attitude data and the second attitude data by filtering a high frequency signal that exceeds the cutoff frequency in the first attitude data and the second attitude data using the frequency domain low-pass filter; generate a target attitude corresponding to the exposure time at which the image acquisition device is located based on the low frequency signal in the first attitude data and the second attitude data that does not exceed the cutoff frequency; and obtain a stabilized target image by stabilizing the image to be stabilized according to the target attitude.
This invention relates to image stabilization in acquisition devices, addressing the challenge of compensating for unwanted motion during image capture. The device includes a processor, an image sensor, and a spatial attitude sensor, which measures the device's orientation. The processor acquires an image frame and its exposure time, then collects attitude data before and after exposure. The first and second sets of attitude data may each consist of one or more measurements. A frequency domain low-pass filter is applied to remove high-frequency signals exceeding a cutoff frequency, retaining only low-frequency signals. These filtered signals from both data sets are used to determine a target attitude at the exposure time, representing the device's stabilized orientation. The original image is then adjusted based on this target attitude to produce a stabilized output image. This approach improves image quality by mitigating motion-induced blur, particularly in handheld or moving devices. The system dynamically compensates for low-frequency motion while suppressing high-frequency noise, ensuring accurate stabilization.
15. The image acquisition device of claim 14 , wherein the processor is further configured to: acquire a number of the frames of the image to be stabilized acquired by the image sensor in the image acquisition device, the image to be stabilized including a time stamp.
The invention relates to image stabilization in image acquisition devices, addressing the problem of motion-induced blur in captured images. The device includes an image sensor that captures frames of an image to be stabilized, each frame containing a time stamp indicating when it was acquired. A processor in the device is configured to analyze these frames to correct for motion artifacts. The processor acquires a specified number of frames from the image sensor, processes the time-stamped data to detect and compensate for motion, and generates a stabilized output image. The stabilization process may involve aligning frames based on their time stamps, interpolating motion paths, or applying digital corrections to reduce blur. The device may also include additional components, such as a motion sensor, to further enhance stabilization accuracy. The invention aims to improve image quality in handheld or moving devices by dynamically adjusting for motion during capture.
16. The image acquisition device of claim 14 , wherein a first time period corresponding to the first attitude data and a second time period corresponding to the second attitude data are the same.
This invention relates to an image acquisition device designed to capture images while compensating for motion or attitude changes. The device includes a sensor for detecting attitude data, such as orientation or movement, and a processor that adjusts image capture parameters based on this data. The device is particularly useful in scenarios where motion stability is critical, such as aerial imaging, handheld photography, or surveillance systems. The device operates by collecting first and second sets of attitude data over the same time period. The processor uses this data to determine adjustments needed to stabilize the captured images. For example, if the device detects a tilt or rotation, it can adjust the camera's lens position, exposure settings, or image processing algorithms to correct for distortion or blurring. The synchronization of the first and second attitude data ensures that the adjustments are applied consistently, improving image quality. The invention addresses the problem of motion-induced image degradation by providing real-time compensation. Traditional systems may rely on post-processing or mechanical stabilization, which can be less effective or introduce latency. By integrating attitude sensing and processing directly into the image acquisition device, this solution offers a more responsive and accurate stabilization method. The device can be implemented in various imaging systems, including cameras, drones, and mobile devices, to enhance image clarity in dynamic environments.
17. The image acquisition device of claim 14 wherein a delay between the exposure time of the target attitude and an exposure time of an actual attitude of the image acquisition device does not exceed a delay threshold.
This invention relates to image acquisition devices, particularly those used in dynamic environments where precise timing and attitude alignment are critical. The problem addressed is the delay between capturing an image at a target attitude (desired orientation) and the actual attitude of the device during exposure, which can lead to misalignment or blurring in the captured image. The image acquisition device includes a sensor for detecting the device's attitude (orientation) and a controller that adjusts the exposure timing based on this data. The controller ensures that the exposure occurs when the device's actual attitude closely matches the target attitude, minimizing any delay between the two. A delay threshold is defined to limit the acceptable time difference between the target and actual exposure attitudes. If the delay exceeds this threshold, the device either adjusts the exposure timing or discards the image to prevent misalignment. The system may also include a stabilization mechanism to reduce physical vibrations or movements that could affect attitude accuracy. The controller continuously monitors the device's attitude and dynamically adjusts exposure parameters to maintain alignment within the delay threshold. This ensures high-quality images even in unstable or rapidly changing environments, such as aerial or underwater imaging. The invention improves accuracy in applications requiring precise attitude alignment, such as surveying, remote sensing, or scientific imaging.
18. The image acquisition device of claim 14 , wherein the processor is further configured to: segment the image to be stabilized based on a segmentation method to obtain a plurality of sub-images; and stitch one or more of the plurality of sub-images based on the target attitude corresponding to the exposure time of the image to be stabilized to obtain a frame of stitched image, the stitched image being the stabilized target image.
This invention relates to image stabilization in image acquisition devices, particularly for correcting motion-induced distortions in captured images. The problem addressed is the instability of images caused by device movement during exposure, which results in blurred or distorted outputs. The solution involves a processor that segments an image to be stabilized into multiple sub-images using a segmentation method. These sub-images are then stitched together based on a target attitude corresponding to the exposure time of the original image, producing a stabilized target image. The segmentation method divides the image into smaller regions, and the stitching process aligns these regions according to the target attitude, which compensates for motion during exposure. The processor ensures that the final stitched image is free from motion artifacts, improving image clarity and quality. This approach is particularly useful in applications where image stability is critical, such as photography, surveillance, and medical imaging. The invention enhances the performance of image acquisition devices by dynamically correcting motion distortions in real-time or post-processing.
Unknown
June 2, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.